Police van equipped with CCTV. (Source: Wikicommons)
It is being reported that biometric automatic facial recognition (AFR) systems used by British police have now resulted in their first arrest. Police in Cardiff, Wales claim to have made the arrest on 31 May, but have provided very little information about it beyond the initial news release. This arrest is significant because – regardless of what the police claim – it clearly marks the shift from the testing of these systems to their operational deployment. In other words, the British public can expect these arrests to happen with increasing frequency for the foreseeable future.
For the past few years, 21WIRE has been watching and reporting on the development of biometric tracking and identification systems. Certain systems are concerned with ‘gait analysis‘ – identifying people by how they walk and move. Researcher and 21WIRE contributor Pippa King focuses on Biometrics in Schools. A year ago, 21WIRE reported on Britain testing ‘millimetre microwave’ systems that reportedly can detect armed individuals in a crowd.
But AFR systems are arguably more sinister than some other types of biometric tracking. Unlike fingerprints or DNA, our faces are usually on display all the time, and can be captured by CCTV systems, cameras on police vehicles, police body cameras, or any other type of camera. As Alastair MacGregor, the UK’s first Commissioner for the Retention and Use of Biometric Material, wrote in 2014:
“… a searchable police database of facial images arguably represents a much greater threat to individual privacy than searchable databases of DNA profiles or fingerprints”.
In addition, AFR systems go beyond the simple identification of individuals walking around in public. AFR may also learn to analyse behaviour to discern mood and emotion. And the US government is currently testing a system that can perform facial recognition through car windows. And some AFR systems are designed to work in real time, tracking and identifying people as they move around, rather than by studying the footage later on.
Identifying an individual by his or her face is fairly easy for a human to do; it is much more complicated for machines. There are many variables, and as a result some facial recognition systems are better than others. For instance, in 2014 it was reported that the DeepFace system used by Facebook outperformed the FBI with its Next Generation Identification (NGI) system. The Verge reported:
“Given a suspect’s face, NGI returns a ranked list of 50 possibilities, and only promises an 85 percent chance of returning the suspect’s name in the list. To put it another way, even when you give NGI 50 guesses, it still lets one in seven suspects off the hook.”
It is now three years later and the system has likely improved, but the point is that although the state may want us to think AFR systems are infallible, they are capable of making mistakes. Faces are not quite as unique as fingerprints, and there is a danger of false identification of suspects.
Photo: Nikolaos S. Karastathis. Source: Wikicommons
And of course then there is private data: CCTV footage owned by shops and businesses. For example, a system called Facewatch serves as a facial recognition version of YouTube; shops and businesses share their footage of shoplifters and problem customers with the entire network of other participating businesses. Though Facewatch itself is not an AFR system on its own, it is compatible with AFR systems that can scan faces in real time. As Ars Technica points out, this could result in people being removed or permanently banned from these businesses, without any way to challenge the decision or prove their innocence.
As a final word on the dangers of AFR, 21WIRE previously warned:
“Far from tracking criminals, the new system may eventually be used to track and monitor individuals who hold what authorities consider to be “anti-government views”. This could range from anyone engaged in a lawsuit against the police or government, to anti-fracking activists, anti-big pharma, anti-GMO, anti-vaccine, parents who home-school their children, or even journalists who hold opinions that differ from the government’s official party line.
“Common sense and history dictates that once a system such as this is in place – no matter how many assurances central government Mandarins might give to symposium panels and media – there is absolutely nothing that can really prevent a centralised police apparatus from both over-using and misusing these Big Data capabilities.”
It’s important to note that the WalesOnline story reposted below reads like a press release from the police. The title of the story refers to the arrest, but only two of the opening sentences in the story mention it at all. The rest of the story comprises information provided by the police about their use of the facial recognition system, including lengthy quotes from two different officers. The article is almost certainly the result of a public relations push by Welsh police; this indicates their desire to acclimatise the public to this new reality, a stepping stone to their goal of gaining the public’s acceptance so that biometric facial recognition systems can be deployed nationwide.
And make no mistake: it will be nationwide. As Chief Superintendent Jon Edwards points out:
“South Wales Police has secured funding from the Home Office to develop automated facial recognition technology for policing.”
The Home Office is a department of the national (not local) government. If the Home Office is funding AFR in South Wales, it is a near certainty that they will soon fund its rollout right across the UK.
More on this story from WalesOnline…
CCTV cameras used for automatic number plate recognition (ANPR) in London. (Source: Wikicommons)
Ruth Mosalski
WalesOnline
The first arrest using new facial recognition software has been made.
South Wales Police has become the first force in the UK to use the equipment.
The first arrest was made on Wednesday but it was not related to the Champions League final.
Real-time cameras linked to facial recognition software will monitor people in and around the city centre.
The images will be use identify people who exist on pre-determined watch lists, usually used for terrorists and hooligans. It will also be used to monitor to ticket touts.
The force has also been given funding for a separate trial of software that enables them to cross reference CCTV images and other picture with their database of 500,000 custody images.
Police vehicles have been spotted around the city labelled as using “facial recognition”.
Continue reading this story at WalesOnline
READ MORE BIOMETRICS NEWS AT: 21st Century Wire Biometrics Files
SUPPORT OUR WORK BY SUBSCRIBING & BECOMING A MEMBER @21WIRE.TV
